While walking past a series of LinkNYC kiosks last week it seemed strange to me that every single one of the tablet devices seemed to have been knocked out of service. These are machines I pass every day and the tablets have been pretty reliable. By the time I passed a third darkened LinkNYC tablet I realized the problem: I was wearing polarized sunglasses, which make LED screens look either very faint or completely turned off. I took off those glasses and voila, there was the great Smart City tablet in all it splendor. I laughed and thought it would be something if sunglasses like this could also make the advertising panels look blank. One could build a cottage industry around such a piece of corrective eyewear.
The sunglasses-making-tablet-screens disappear thing reminded me I’d been meaning to revisit the map I put together a couple of weeks ago. The map is intended to show real-time functional status of the LinkNYC kiosks’ WiFi, tablet, and phone call features. This map uses data from the NYC.gov data portal. The data is provided by CityBridge, which claims it provides “the most current listing of LinkNYC Kiosks, their location, and the status of the Link’s wifi, tablet, and phone.”
After a first glance at the dataset a couple of months ago I found numerous errors but decided to give it another try, thinking maybe kinks needed to be worked out of the data, which is said to be 100% automated (a finely insidious Smart City buzzword).
In my first attempt at re-verifying the accuracy of this data I was happily surprised. A kiosk location in which the tablet was not working was accurately reported in the dataset as not working. The next day the tablet was working again, and the dataset seemed to reflect this, reporting the tablet was up. It felt pretty cool, as if the data had a pulse. It had me thinking that maybe this Smart City thing isn’t as stupid as I thought.
Sorry to say it but since then the percentage of accurate status reports of the tablets and phones has been very low, making the one accurate report earlier look like a random fluke. In the last 2 weeks I’ve encountered about 20 to 25 kiosks with non-working tablets or non-working telephone calls. In all but or two or three of those instances CityBridge’s dataset reported the tablet devices as up and running, and in virtually all cases it reported that the phone call feature was working when it was not. In some cases the phone call feature works when the tablet screen does not but those are few and far between. I have not checked on the working status of the WiFi and I do not intend to, but I did do a spot check of about 20 kiosks’ WiFi some months ago and found the speeds and availability to be all over the place, some connections blazing fast, others not so much, and still others that never connected at all. The dataset from CityBridge makes it look like net network’s WiFi uptime is close to 100%, a claim I find very hard to believe.
The dataset was originally supposed to be updated hourly, but at this point is it not clear to me how frequently it is refreshed. I download the dataset once an hour. I had spotted a tablet near me that stopped working, having worked fine the day before. I looked that up on my map and found it was still reported as up and running. To be generous I gave CityBridge 24 hours to update the status of this device. It made no difference. CityBridge reported that tablet as up and running 72 hours later, but it was still not working.
To be clear, I do not expect these machines to be 100% up and running, or even close to it. I am not complaining about some tablets being offline, rather I’m questioning why CityBridge’s dataset is even being published in the first place.
Had the data been more reliable I might have expanded the mapping I did to include three separate maps, just for clarity, each showing yea or nay icons for individual functions. The present map of LinkNYC Kiosk Status is a little unwieldy but it works, not that anyone should rely on it for accuracy. I’ll expand on it to include pages with all available information about the LinkNYC kiosks, along with comment boards where you can report on the accuracy of the status as reported by CityBridge, or just discuss a little bit the interesting experiences you’ve had involving these individual machines.
I’ve been giving no shortage of my attention to the LinkNYC machines of late, but I’ve demurred from writing about it for what I think is a simple and understandable reason: Not very many people really care about these things. A lot of New Yorkers have never even encountered a LinkNYC kiosk, and among my circles I find that most who have seen them do not even know what they are until I explain it to them. These machines are deposited onto the curb as if their purpose is self-evident. An “About” screen is found on the tablet, but it lurks below the fold where, to my non-polarized eyes, the content can be hard to read, especially during daylight hours.
What has never failed to surprise me is how much bad content floats through these things. Lately I’m starting to think that maybe it does not actually matter, that we should accept these things for what they are, which is a bit of self-resigned apathy that further intrudes on my desire to write more about this subject.
By “bad content” I refer to the appearance of old AP headlines from weeks or even months previous, weather snapshots that have no connection to the current reality, phone calls that disconnect after 30 seconds (or never connect at all), and so on. The weather content has proven to be entirely unreliable. On Saturday, while the temperature was cruising at about 73°, every single tablet screen I saw was locked in at 59°. The bigger screens seemed closer to reality, reporting a temperature in the low- to mid-70s. But I’ve seen those big screens report current temperatures as far off as 25° Fahrenheit.
Something of note regarding the weather snapshot on the tablet screen: CityBridge recently included what is (allegedly) the current temperature in Celsius. Until a week or so ago only Fahrenheit was presented, and Fahrenheit remains the only unit of measurement seen on the larger advertising screens.
This little flourish of adding Celsius to the tablet screen probably captured the attention of absolutely nobody, but I saw it as a potentially positive sign that someone at the company might have been taken on the task of improving the weather product on the tablet. Maybe this signaled that someone would actually do some QA on it and figure out why so many of the devices display information that is completely wrong.
Alas, from my observations this week I’ve arrived at the disappointing conclusion that adding the current temperature in Celsius was simply a well-intentioned gesture to make the temperature more readable for tourists, since most of the world uses Celsius on its thermometers. As for the integrity of the data itself with LinkNYC the seemingly simple act of obtaining the current weather conditions turns out to be like a multiple choice quiz, or a game of “Guess The Current Temperature”. A few nights ago I passed three machines in a row. The first one said 71°, the next one said 79°, another said 82°. Weather snapshots also appear on the large advertising screen. In the case of the machine where the tablet said it was 82° the large screen reported 71°. While Google might be able to change neighborhood names by automated algorithmic fiat at least we will never arrive at a point where its inaccurate weather reports will actually change the temperature around us.
The image above, which looks like it was taken at 5 in the morning, was actually shot at 2pm, when it was about 73°. I spotted at least a dozen tablet screens that reported similarly wrong weather conditions.
But stale weather content is not the only bothersome glitch about these machines. A few days ago I spotted weeks-old headlines about Tropical Storm Gordon approaching the Gulf Coast, followed by similarly outdated warnings that Hurricane Florence was approaching the Carolinas. Other headlines of similarly questionable timeliness came and went but in those cases I was not familiar with the subject matter enough to say if it was dated or not.
You can argue, and I would not disagree, that people are not stupid. Most of us who see something like a hurricane warning for a storm that came and went weeks ago will, depending on how informed we are on current events, know that this content is in error. But what kind of “War Of The Worlds” scenario might emerge when these LinkNYC kiosks start blasting out emergency alerts for disasters that are not happening, or Amber Alerts from 3 months ago? It’s so amateurish the way some of these LinkNYC content products are being managed, and no one should assume it is because publishing this type of content is as difficult as CityBridge makes it look.
Some would say that these little inaccuracies are inconsequential growing pains of the Smart City. I disagree. This product should be considered well out of its Beta stages now. If it is 100° out but LinkNYC can, without consequence, get away with telling us it is 72° then it seems inevitable that content quality control at the company will get more and more lax and these seemingly little errors will be usurped by more significant ones.
For now, at least, they’ve shown me some mercy by getting rid of the profoundly useless bus arrival time screens, which I tried so hard to make useful but just could not make it happen. The bus arrival screens are just the type of thing Smart City groupies glom on to, pointing at its alleged usefulness without actually attempting to interpret it or make use of it in the spirit it appears to be intended. The LinkNYC bus arrival screens contained nothing more than random and untreated blasts of algorithmic effluvia. What the bus arrival time needed is what a lot of LinkNYC products need: human intervention. How hard a job would it be for someone to go in person to kiosk locations or even to just look at a map to determine which bus lines are most relevant to the area? How hard is it for the data feeding the weather content to be centralized? How hard is it to not post AP headlines from weeks ago?
